SUAV:Q - A Hybrid Approach To Solar-Powered Flight

Ruben D’Sa, Devon Jenson, and Nikolaos Papanikolopoulosdsaxx005@umn.edu, jens1172@umn.edu, npapas@cs.umn.edu

Department of Computer Science, University of Minnesota

Abstract— Selecting an aerial platform for an applicationtypically requires compromise. A choice must be made betweenthe flight time and long-range capabilities of a fixed-wingaircraft or the maneuverability and stationary characteristicsof a multi-rotor platform. Recent developments of small-scalesolar-powered UAVs have leveraged the advances in solar cell,energy storage, and propulsion system technology to reachextended flight times capable of all-day and multi-day flight.

This paper presents the concept of a small-scale hybridunmanned aerial vehicle capable of augmenting the maneu-verability of a quad-rotor with the energy collection andsupply of a solar-powered fixed-wing aircraft. An investigationinto the aircraft design, transforming mechanism, and energymanagement of the multi-state system is presented.

A proof-of-concept prototype has been constructed to demon-strate the airframe operating in a quad-rotor configuration.Power electronics capable of simultaneous battery chargingand power loading from a solar array have been validated.Additional work in optimization of the propulsion system andairframe needs to be completed to maximize the performanceof the hybrid system.

I. INTRODUCTION

In recent years, aerial robotics has seen tremendous growthwithin the areas of both fixed-wing and multi-rotor flight.Although great achievements have been made in the domainof aircraft control, an area that has seen little developmentis that of solar-powered flight endurance and performance.Recently, a number of groups have shown development inthe design of small-scale solar-powered aircraft [1], [2],[3]. Thanks to advancements in motor controller and motorperformance, solar cell efficiency, and battery density, moresolar energy can be captured and stored than what is requiredto fly at level flight. As a result, a number of these systemshave demonstrated day-long and multi-day flight capability.

Several fixed-wing and flying wing systems have beencapable of solar-powered flight [4], [5], [6]. To achieve anumber of design goals, the size of these aircraft range be-tween 4 meters [3], [7], [1] and 5.8 meters [2]. These systemshave relatively high aspect ratio wings and, in order to meetthe multi-day flight design goals, have correspondingly longwingspans. One of the greatest challenges with fixed-wingaircraft, compounded by a low Reynolds number and highaspect ratio wings, is maneuverability.

Multi-rotor systems are commonly used in applicationsthat require both high maneuverability and the ability to holda fixed spatial position. Their applications range from identi-fication in search-and-rescue to characterization of nitrogendeficiencies in corn fields [8]. However, maneuverabilityand control come at the cost of high power consumption,

Fig. 1: Fixed-wing and quad-rotor states of the proposed SUAV:Q hybrid aircraftdesign.

resulting in short flight times. This is in contrast with thehigh efficiency, long-flight capable fixed-wing systems.

As a solution to the restrictions faced when selecting eitherplatform, this paper presents the design of a re-configurablesolar UAV that is capable of transforming between fixed-wing and quad-rotor states. A platform capable of both statesremoves individual limitations and combines the strengthsof both systems. In quad-rotor state, the aircraft cannotsupply enough energy from solar power alone and will haveto rely on stored energy. Once stored energy is close tobeing depleted, the aircraft will transition into a fixed-wingstate where the on-board batteries will be able to recharge,allowing the process to repeat. A related approach to energymanagement has been applied to a solar-powered seaplane[9], where the ability to land on water enables long-termautonomous flight. Similarly, in this proposed design, verticaltakeoff and landing allow the aircraft to wait on the groundfor suitable solar conditions and land in the absence of them.By combining this functionality with a solar-powered system,the ability to achieve multi-day operation becomes feasibleat very small scales. Traditionally, the size of solar-poweredUAVs were limited due to energy storage requirements as theaircraft would have to store enough energy to supply powerthrough the night. By relieving the constraint of carryinga large battery and replacing it with vertical takeoff andlanding, the size of the aircraft can be dramatically reduced.

II. ROBOT DESIGN

This section discusses several important considerations inthe design of the robotic aircraft: aircraft stability and weight

distribution, the mechanical transition between flight modes,and the transforming linkage design and analysis.

A. Approach

A number of propeller-based fixed-wing aircraft capableof vertical takeoff and landing have been developed. Thesesystems range from very large-scale aircraft, such as theBoeing V22 Osprey, to very small-scale systems, such as theAeroVironment SkyTote [10]. In the case of the SkyTote, thedesign focused on high speed fixed-wing flight as opposedto the reduced energy consumption associated with lowspeed flight. While vertical take-off and landing sailplaneshave been pursued [11], [12], this approach requires a largesurface area while in a hovering state. Sailplanes also havelimited hovering robustness due to their aircraft geometry.To develop a system that combines a high aspect ratio wingwith the compactness of a quad-rotor, a sectioned reflexedchord plank wing is used. While control of a flying wingis more involved, the transformable design lends itself to avery stable quad-rotor design.

B. Aircraft Design

To facilitate the transitions between fixed-wing to quad-rotor states, a plank flying wing geometry was pursued.This design decision allows for motor and propulsion systemassemblies to be mounted at the same longitudinal position,as opposed to a staggered arrangement. To provide rollstability, active hinges are used to transition the aircraftbetween wing states and are designed to have a hard stopthat creates a built-in dihedral angle as shown in Figure 2.

Fig. 2: Front view of the hinge section transitioning between fixed-wing and quad-rotorstates. θd is the dihedral angle. ωin and ωout represent the input and output angularvelocity respectively. rin is the length of the servo arm and rout is the length betweenthe ground pivot and wing section weight vector which is assumed to be constant.

One of the challenges in designing aircraft with a flyingwing geometry is lateral stability. In the selection of anaerofoil for a traditionally designed fixed-wing sailplane, theaircraft’s tail elevator is used to counter the pitching momentof the main wing and center of gravity. In the case of aflying wing, no tail exists to counter this moment and boththe aerofoil design and aircraft weight distribution must beselected accordingly. As discussed in [13], the elevator ona traditional aircraft can be designed into a single wing,resulting in a reflexed chord line aerofoil. By using a reflexedchord line aerofoil and adjusting the center of gravity of the

aircraft to be in front of the neutral point (nose heavy), thereflexed tail section of the aerofoil will counter the pitchingmoment induced by the center of gravity and create a stableequilibrium point. The MH49 aerofoil was selected due toits performance at low speeds. Components were positionedto place the center of gravity in front of the neutral point asshown in Figure 3. A list of component weights are given inTable I. As the distance between the neutral point and centerof gravity increases, the pitch stability increases, requiring alarger control surface. In order to minimize the area of thewing that is unable to be covered by solar cells, the centerof gravity was adjusted to minimize the necessary controlsurface area while still providing adequate control authority.

Fig. 3: Side view of the wing section illustrating the center of gravity (C.G.) andneutral point (N.P.) locations on a single wing section.

Component IndividualSection (g)

EntireSystem (g)

Solar Cells (8x SunPower C60) 108 430Maximum Power Point Tracker 90 360Battery Management System 31 124Battery Mass (6x NCR18650B) 276 1104Airframe 332 864Propulsion System 94 376Camera and Sensors N.A. 0 to 300Total 715 3364 to 3664

TABLE I: Estimated mass of the solar UAV prototype components.

In order to find the operating points that maximize energyperformance of the aircraft, a vertex lattice method (VLM)analysis was performed using the software package XFLR5[14]. An iterative approach was taken between adjustingcomponent placement and the resulting global pitching mo-ment. VLM analysis was performed using fixed lift polarswith the lift equated to the mass of the aircraft. The resultsare shown in Figure 4. Using Figure 4(b), the optimum flightvelocity for a system mass of 3.2kg is identified at 14.25m/swhich can then be used to find the angle of attack in Figure4(a).

Velocity [m/s]

(a)

0 10 20 30 40 50

Alp

ha (

De

gre

es)

0

1

2

3

4

5

6

7

Velocity [m/s]

(b)

0 10 20 30 40 50

Cl/C

d

0

5

10

15

20

25

2kg

3kg

4kg

5kg

6kg

3.2kg

Fig. 4: Aircraft performance analysis using XFLR5 with VLM fixed lift polars equatedto the mass of the aircraft. (a) Minimum angle of attack at a given velocity to ensurelift. (b) Ratio of coefficient of lift to coefficient of drag as a function of velocity.Maximum performance for a system mass of 3.2kg occurs at 14.25m/s with a Cl/Cdratio of 23.40.

Fig. 5: Transition path from quad-rotor to fixed-wing states.

C. Forced State Transition

In order to minimize the size and mass of actuators used,the flight trajectory upon transition is selected to maneuverthe aircraft into a state that minimizes the hinge actuationtorque. By flying the aircraft in a near vertical state, thegravitational forces acting on the wing segments are orientednearly parallel to the axis of hinge rotation, minimizing thetorque required for actuation. A demonstration of this tran-sition is shown in Figure 5. While transitioning from fixed-wing to quad-rotor state, the air resistance of the wing willresist transition. However, by minimizing the translationalvelocity of the aircraft, this force is reduced as shown inFigure 6. Currently, hinge operation is teleoperated and trig-gered by the flight controller. With additional modificationsto flight controller firmware, on-board inertial measurementcan be used as feedback to insure robustness of automatedtransitions. Management of the propulsion system torque inboth states are shown in Figure 7.

D. Linkage Design

To provide actuation of the individual wing segments,a four-bar hinge mechanism is used. Shown in Figure 2is the linkage in both fixed-wing and quad-rotor states. Inorder to minimize actuator energy power consumption in thefixed-wing state, the hinge enclosure has a built-in hard stopthat limits the dihedral angle to the prescribed 8.7 degrees.A servo actuator was chosen over a linear actuator for itsspeed of actuation at the consequence of lower overall outputtorque. The 4-bar linkage was graphically synthesized andits mechanical advantage was evaluated using Equation (1)from [15] and plotted as a function of the angular positionof the servo link in Figure 8. One of the mechanism designobjectives was to build the system into the wing sectionsand minimize the gap surface between wing sections. Thiswas achieved by anchoring the wing segment to the outputlink as opposed to the coupler link. Input no-load angularvelocity is assumed to be a constant 428deg/s from a PowerHD 1501MG with peak torque rated at 17kg·cm.

M.A.=rin

rout

ωin

ωout(1)

ωout =rin

routωin

sin(θ3 −θ2)

sin(θ3 −θ4)(2)

Fig. 6: Side view of forces and torques acting on the active hinge during transitionbetween fixed-wing and quad-rotor states.

(a)

(b)

(c)Fig. 7: (a) To manage the torque produced by each propeller and motor combination,the direction of rotation alternates such that the net induced torque by the propulsionsystem is balanced in both fixed-wing and quad-rotor configurations. (b) Each of thewing sections contain identical power electronics and battery hardware in order tomanage the weight distribution across states. Shown in (c) is a comparison betweenthe surface area difference between fixed-wing and quad rotor states.

Driver link angle [Degrees]

(a)

200 250 300Me

ch

an

ica

l a

dva

nta

ge

(F

out/F

in)

0

0.1

0.2

0.3

0.4

Driver link angle [Degrees]

(b)

200 250 300

An

gu

lar

Ve

locity [

de

g/s

]

-0.2

0

0.2

0.4

Coupler

Output

Driver link angle [Degrees]

(c)

200 250 300

An

gu

lar

Po

sitio

n [

de

g]

-150

-100

-50

0

Coupler

Output

Fig. 8: Mechanical advantage and linkage angular velocity and position plotted as afunction of driver link position.

III. ENERGY FRAMEWORK

This section establishes a basic framework for evaluatingthe performance of a transformable solar-powered aircraft.A generalized system model is defined and used to analyzethe performance of fixed-wing and rotor states for differentbattery sizes and payload. This analysis is concerned withhybrid operation of the aircraft, i.e. frequent transitionsbetween fixed-wing and rotor states, and focuses on daytimeflight where sufficient energy is available for rotor flight.

A. Hybrid System Model

The hybrid model consists of three states: fixed-wing,rotor, and ground. Fixed-wing and rotor states correspondto the right and left side of Figure 5. Their differences in liftgeneration provide a trade-off between power consumptionand maneuverability. In ground state, the aircraft chargeson the ground, typically in a fixed-wing orientation tomaximize solar power intake. The following equations formthe basis of the model. In level flight fixed-wing state, powerconsumption is given by [2] [7]

Pfixed(mtotal) =Plevel

ηprop=

1ηprop

· CD

C32L

·

√2(mtotal ·g)3

ρ ·Awing(3)

where ηprop is the propulsion system efficiency, CD andCL are the drag and lift coefficients that minimize levelflight power, mtotal is the total system mass (including energystorage and payload mass), g is the acceleration of gravity,ρ is the air density at a constant altitude, and Awing is thewing reference area.

In rotor state, the power consumed during hover conditionsis given by

Protor(mtotal) =Crotor ·mtotal32 (4)

where Crotor is a constant defined for a particular rotortopology and mtotal is the total system mass.

For solar power intake, the aircraft is assumed to beoriented with its wing segments open (as in fixed-wing state).This is shown in Figure 9.

Fig. 9: Dihedral angles φ1 and φ2 of the SUAV:Q airframe in fixed wing state.

The solar power intake is given by

Psolar(t) = max[0, I ·APV ·D ·ηPV · sin(π · t

tday)]

(5)

D =cos(φ1)+ cos(φ2)

2(6)

where I is the peak solar irradiance, APV is the solar panelarea, D is a constant that considers the angle of incidence

due to the dihedrals, ηPV is the solar panel efficiency, andtday is the length of day.

The available power in each of the three states is given by

Pavail, gnd(t) = Psolar(t) (7)

Pavail, fixed(t) = Psolar(t)−Pfixed(mtotal) (8)

Pavail, rotor(t) =−Protor(mtotal) (9)

where it is assumed as a worst case analysis that no solarpower is available in rotor state. Available power can bedefined as the excess solar power in a given state. For it tobe utilized, it must be able to be stored on-board for lateruse. In a deficit of solar power, power must be supplied fromon-board storage to remain in the current state.

B. Power Levels and Energy Management

The three different states of the proposed hybrid vehicleprovide different levels of available power. Energy can bemanaged by transitioning between these states to ensure amaximum use of available power and a minimum level ofrobustness. From an energy perspective, the aircraft movesbetween higher and lower energy states. For example, rotorflight can be thought of as a high energy state becauseit requires stored energy and can only be maintained fora relatively short period of time. Transitioning to a lowerstate (i.e., rotor to fixed-wing or rotor to ground) relaxes theenergy required by the system.

Maximum use of available power is achieved by avoidingsaturation of on-board energy storage. For hybrid aircraft,this means transitioning to rotor state before maximumcapacity is reached. The battery’s state of charge can alsolimit the charge rate and, as a result, limit the use of availablepower. It may be beneficial to set an upper energy thresholdto ensure a minimum charge rate at all times.

A metric for robustness is stored on-board energy. In theevent that available power is negative, due to flight power re-quirements or worsening solar conditions, stored energy canbe used to transition to a lower state or maintain the currentstate. Similar to an upper threshold, the system can maintaina minimum amount of stored energy by transitioning to alower state if the corresponding minimum energy thresholdis reached. This guarantees a minimum level of robustnessin all states, independent of the physical parameters of thesystem.

C. State Performance Parameters

An important parameter for hybrid aircraft is the chargeor discharge time to a specified upper or lower threshold,Ebat, thresh, in a given state. This is given by the time t thatsolves the following equation

t∫t0

Pavail,state(t) dt +Ebat(t0) = Ebat, thresh (10)

As the fixed-wing state is used for both flight and intake ofavailable energy, it has an associated charge time tfixed. Rotorstate requires stored energy and has an associated discharge

time trotor between thresholds. An upper threshold can beassigned as the energy required for an application specificrotor operation and, as described in the previous subsection,a lower threshold can be used to ensure a minimum level ofrobustness.

Because charge and discharge times depend on solarconditions and dynamic power consumption, a more generalparameter is the ratio of time spent in a state,

tratio, state =tstate

ttotal(11)

where ttotal is the total time spent in ground, fixed-wing, androtor states.

A majority of hybrid operation occurs during the periodwhen the fixed-wing state has available power (labeled tavail).In this length of time, fixed-wing level flight is completelypowered by solar and available power is stored for lateruse in rotor state. To optimize for both available energyand hybrid flight time, the aircraft should begin flight whenPfixed is first equivalent to Psolar (assuming on-board storagedoes not saturate before this time) and land after time tavail.At the expense of stored energy, flight time can be extendedbefore and after these points. These times and tavail are givenby,

t01 =tday

πarcsin

( Pfixed

I ·APV ·ηPV

)(12)

t02 = tday −2t01 (13)

tavail = t02 − t01. (14)

D. Simulations - Battery Size and Payload

This subsection explores the effect of battery size andpayload on state performance using the above system model.The conceptual design parameters used in the simulations areshown below in Table II. An estimate on the rotor constant,Crotor, for the prototype design was determined empirically(see Section V).

ηprop 0.55ρ 1.225 kg/m3

Awing 0.644 m2

Crotor 50.41 W/kg3/2

I 1000 W/m2

ηPV 0.22Awing 0.644 m2

φ1 8.7 degreesφ2 17.4 degrees

TABLE II: SUAV:Q conceptual design parameters used in simulations.

Fixed-wing performance is dependent on level flight powerconsumption. Figure 10 compares the available power andenergy over the course of a 12-hour day with 30W and60W fixed-wing flight powers. In each instance, the aircraftremains in ground state until Psolar is greater than Pfixed tooptimize for both storage of available energy and flight time.It can be seen that as Pfixed increases, both Pavail,fixed andtavail decrease. This results in longer fixed-wing charge timesand less overall available energy for larger payloads andbatteries. If an upper energy threshold is required for rotor

state operations, these operations will be less frequent. Figure11 shows total available flight time (tavail) and fixed-wingavailable energy (Eavail, fixed) as a function of total systemmass. For each mtotal, the coefficient of lift and the coefficientof drag that minimize level flight power were used. It canbe seen that using a larger battery to increase robustnessthresholds or extend flights beyond tavail will degrade fixed-wing performance. Increasing battery size results in anapproximately linear decrease in available energy.

Po

we

r [W

]

0

50

100

150

Psolar

Pfixed

Po

we

r [W

]

0

50

100

150

Pavail

Time [h]

0 6 12

En

erg

y [

Wh

]0

400

800

Eavail

Fig. 10: Comparison of two different fixed-wing flight powers and their effects onavailable power and energy over the course of a 12-hour day; dashed lines represent60W and solid lines represent 30W level flight powers.

mtotal

[kg]

0 1 2 3 4 5 6 7

t avail [

h]

0

6

12

mtotal

[kg]

0 1 2 3 4 5 6 7

En

erg

y [

Wh

]

0

550

1100

Eavail

Esolar

Efixed

Fig. 11: Available flight time, solar energy, fixed-wing energy, and available energyas functions of total system mass.

Because rotor flight requires stored power, battery size isa critical parameter for rotor state performance. Figure 12shows the effect of battery size on rotor flight time for variousaircraft masses (i.e., mass of aircraft excluding battery mass).In this analysis, Protor was assumed constant. The transitionbetween fixed-wing and rotor state is governed by the upperand lower energy storage thresholds. In this simulation thethresholds were set to the minimum and maximum energycapacity of the battery. For the simulated system mass andpayloads, increasing battery size increases rotor flight timewith diminishing returns. This means a small increase inrotor flight time requires a longer fixed-wing charge time.In other words, the proportion of total flight in rotor statebecomes smaller with increasing battery size. Figure 13shows the state of each aircraft from Figure 12, at theoperating point of twelve battery cells, over the course ofa 12-hour day. It can be seen that increasing maircraft for a

given battery size decreases ttotal, tratio, rotor, and trotor. For bestperformance, the battery should be sized for the maximummaircraft and trotor required by a specific application.

# of NCR18650B Battery Cells

12 24 36 48

t roto

r, m

ax [h]

0

0.5

1

1.5

2

2.5m

aircraft = 1.00kg

maircraft

= 2.00kg

maircraft

= 4.00kg

maircraft

= 8.00kg

maircraft

= 16.00kg

Fig. 12: Rotor flight time (trotor) as a function of NCR18650B battery cells (interpolatedbetween discrete points); mtotal = maircraft +mbattery.

012

tratio

= 0.522

ttotal

= 12.688

012

tratio

= 0.382

ttotal

= 12.039

012

tratio

= 0.183

ttotal

= 12.074

012

tratio

= 0.085

ttotal

= 11.738

Time [h]0 6 12

Sta

te

012

tratio

= 0.029

ttotal

= 11.281

Fig. 13: Aircraft state (0 = ground, 1 = fixed-wing, 2 = rotor) over a 12-hour daywith 12 NCR18650B cells and various aircraft masses. The aircraft masses follow thelegend of Figure 12. tratio is the ratio of total flight time (ttotal) spent in rotor state. Foreach aircraft, once the corresponding t02 was reached, the aircraft transitioned to rotorstate until the remaining stored energy was depleted.

IV. ELECTRICAL HARDWARE

Figure 14 is an illustration of the hardware topology of theSUAV:Q. Each of the four wing sections contain an identicalconfiguration of power electronics and batteries. Within eachwing segment, the placement of components is such that thecenter of gravity lies at the longitudinal mid-section. Theintent of even weight distribution is to maximize stability inboth quad-rotor and fixed-wing states.

Each wing section contains 8x SunPower C60 solar cellscapable of 24 watts. In fixed-wing configuration and withideal solar conditions, the power system is capable ofsupplying 96 watts from the solar array. Due to the I-V characteristics of solar cells, the impedance of the loadmust be closely matched with the output impedance of thesolar array. Given the varied angle of solar irradiance hittingeach angled section of the aircraft, each section requiresa maximum power point tracker (MPPT). Each MPPT isused to track and adjust the voltage operating point ofits corresponding panel to ultimately maximize the amountof solar power available to the aircraft. In addition to theMPPT, each of the lithium ion cells is monitored by abattery protection system to protect against over-voltage andunbalanced cells.

Power is monitored throughout the system using the TIINA219 current shunt and power monitor. Each module

communicates with an ATmega328 over an I2C bus. Voltageand current measurements received by the ATmega328 arestored locally to an SD card as well as relayed in real-timeusing 915MHz radios to a ground station.

Fig. 14: Electrical power and communication topology for the SUAV:Q. Identicalbatteries, power conversion, speed controllers, and actuation servos are used across allfour segments.

To control the aircarft, two Pixhawk autopilots are used.A 4-channel PWM multiplexer is used to switch throttlecontrol of the four electronic speed controllers between thetwo flight controllers, where one is configured for fixed-wing and the other quad-rotor flight. Teleoperation of thesystem is performed using a Spektrum DSMX transmitterand receiver combination. Using an external u-blox LEA-6H GPS module, the system can be set to fly to prescribedGPS way points. A single PWM signal is used to controlall four hinge servos and is configured in the Pixhawk to beremotely controlled by a toggle switch on a Spektrum DX8transmitter.

Time [s]

0 100 200 300 400 500 600 700

Po

we

r [W

]

-10

0

10

20

30

40

50

60

70

80

Battery Input

BMS Output

Resistive Load

MPPT Output

Fig. 15: Solar power system charging 36 Panasonic NCR18650b cells connected in a6s6p configuration. A 10ohm resistive load was connected to the output to representpower draw from the SUAV:Q propulsion system. Initial peak in solar power was dueto array alignment with respect to the sun. Average collected solar power was 68.92W.Test data was collected on Sept. 13, 2015 at Nils Hasslemo Hall, Minneapolis, MN at5:45pm.

V. EXPERIMENTAL TESTS

Figure 15 shows the results from an outdoor test of 32SunPower C60 cells along with a 6s battery managementsystem and 36 Panasonic NCR18650B cells wired in a6s6p configuration. An LT8490 based MPPT was connectedbetween the solar cells and the battery. The battery cells wereprotected by a PCM-LI22.2V20A battery management sys-tem for 6s battery configurations. In order to simulate a step

0 10 20 30 40 50 60 70

Po

wer

[W]

0

200

400

0 10 20 30 40 50 60 70

Voltage

[V

]

14

15

16

Time [s]

0 10 20 30 40 50 60 70

Cu

rrent [A

]

0

20

40

Fig. 16: Battery voltage, current, and power draw from the prototype airframeoperating in the quad rotor configuration. Average power consumption during hoverwas measured at 288.57W with a 4s1p 4000mAh LiPo configuration and a systemmass of 3.2kg. Test data was collected in-doors with minimal wind disturbance.

load response, a 10Ω 200W power resistor was connectedin parallel with the MPPT output. Upon connecting theresistive load, the MPPT output stayed constant, reducing thebattery charging power level to satisfy the load requirementsas expected. Power consumption data and photos of theprototype in quad-rotor configuration are shown in Figure16 and Figure 17.

VI. CONCLUSIONS AND FUTURE WORK

Presented in this paper are the aircraft design and energyframework for a hybrid quad-rotor to solar-powered fixed-wing UAV. Lift characteristics of the aircraft design weresimulated and optimum operating conditions were deter-mined. Performance of the active hinge was evaluated usingtechniques from linkage analysis. Models for the operatingtime across the multi-state system were parameterized to takeinto account temporal changes in solar irradiance and systemmass, illustrating the flexibility of the hybrid system to max-imize use of available energy. Power system electronics werevalidated using 32 SunPower C60 cells and functionality inthe quad-rotor state was demonstrated.

Work is underway to improve the current prototype, mostnotably in the areas of washout aerofoil design and optimalcontrol surface sizing. Validation of prototype operationin fixed-wing and transition between states needs to beperformed. Improvements to the propulsion system will needto be made; the propeller and motor combination will needto be sized to a dynamic thrust rather than a static thrust tominimize level flight power consumption. The general energymodel is currently being expanded to include state transitionenergy and a detailed propeller-motor analysis for fixed-wingand rotor states.

VII. ACKNOWLEDGEMENTS

This material is based upon work supported by the Na-tional Science Foundation through grants IIP-0934327, IIS-1017344, IIP-1332133, IIS-1427014, IIP-1432957, OISE-1551059, CNS-1514626, CNS-1531330, and CNS-1544887.Ruben D’Sa was supported by a National Science FoundationGraduate Research Fellowship No. 00039202.

Fig. 17: Prototype system airframe in both fixed wing and quad rotor configurations.Active hinges were fabricated using an FDM 3D printer and PLA filiment.

REFERENCES

[1] A. Noth, W. Engel, and R. Siegwart, “Design of an ultra-lightweightautonomous solar airplane for continuous flight,” Field and ServiceRobotics, vol. 25, pp. 441–452, 2006.

[2] P. Oettershagen, A. Melzer, T. Mantel, K. Rudin, R. Lotz, D. Sieben-mann, S. Leutenegger, K. Alexis, and R. Siegwart, “A solar-poweredhand-launchable uav for low-altitude multi-day continuous flight,” inIEEE International Conference on Robotics and Automation (ICRA),May 2015, pp. 3986–3993.

[3] S. Morton, R. D’Sa, and N. Papanikolopoulos, “Solar powered uav:Design and experiments,” IEEE/RSJ International Conference onIntelligent Robots and Systems (IROS), Hamburg, vol. 14, no. 1, pp.2460–2466, 2015.

[4] T. Noll, J. Brown, M. Perez-Davis, S. Ishmael, G. Tiffany, andM. Gaier, “Investigation of the Helios prototype aircraft mishap report,NASA,” 2004.

[5] J. Amos, “‘Eternal plane’ returns to Earth. BBC News, 2010. “http://www.bbc.co.uk/news/science-environment-10733998.”

[6] K. Flittie and B. Curtin, “Pathfinder solar-powered aircraft flightperformance,” vol. 4446. AIAA, 1998.

[7] S. Morton, L. Scharber, and N. Papanikolopoulos, “Solar poweredunmanned aerial vehicle for continuous flight: Conceptual overviewand optimization,” in IEEE International Conference on Robotics andAutomation (ICRA), 2013, pp. 766–771.

[8] D. Zermas, D. Teng, P. Stanitsas, M. Bazakos, V. Morellas, D. Mulla,and N. Papanikolopoulos, “Automation solutions for the evaluation ofplant health in corn fields,” IEEE/RSJ International Conference onIntelligent Robots and Systems (IROS), Hamburg, 2015.

[9] R. D. Eubank, “Autonomous flight, fault, and energy management ofthe flying fish solar-powered seaplane.” Ph.D. dissertation, Universityof Michigan, 2012.

[10] A. Koehl, H. Rafaralahy, M. Boutayeb, and B. Martinez, “Aerody-namic modelling and experimental identification of a coaxial-rotoruav,” Journal of Intelligent and Robotic Systems, vol. 68, no. 1, pp.53–68, 2012.

[11] S. Verling and J. Zilly, “Modeling and control of a VTOL glider,”Bachelor Thesis, Autonomous Systems Lab, ETH Zurich, April 2013.

[12] T. Matsumoto, K. Kita, R. Suzuki, A. Oosedo, K. Go, Y. Hoshino,A. Konno, and M. Uchiyama, “A hovering control strategy for a tail-sitter vtol uav that increases stability against large disturbance,” inIEEE International Conference on Robotics and Automation (ICRA),May 2010, pp. 54–59.

[13] M. Hepperle, “Design of flying wing models,” http://www.mh-aerotools.de/airfoils/flywing1.htm, 2015.

[14] A. Deperrois, “Stability analysis using XFLR5,” http://www.xflr5.com/docs/XFLR5 and Stability analysis.pdf, 2010.

[15] A. Erdman, Mechanism Design: Analysis and Synthesis. UpperSaddle River, NJ: Prentice Hall, 2001.